Cryoprotectant perfusion failure modes

Epistemic status: M22 composition is documented (C1). The specific claim "CPAs cannot perfuse closed capillaries" is obvious from first principles but hard to find as a headline claim in published literature — it's treated as a background assumption in the cryobiology community rather than a formal result (C3). The comparative quality between Sparks-style and Nectome-style aldehyde preservation is mostly anecdotal / internal; no head-to-head published EM comparison exists that I could find (C4/C5).

1. What M22 and VM-1 actually are

M22 (21st Century Medicine)

Full composition (Biostasis; cryonics.miraheze.org/wiki/M22; Fahy et al. US patent US8679735B2):

Total ~9.3 M, ~64.8% w/v. Designed for vitrification of whole rabbit kidneys. Slow perfusion is required because high-concentration CPAs are viscous and cells only tolerate them at the terminal concentration for limited time.

Ice-blockers: X-1000 and Z-1000 are proprietary polymers mimicking antifreeze proteins; small additions (1–2% w/v) dramatically suppress ice nucleation.

Noted side-effect: "M22 produces substantial brain shrinking during perfusion" (Biostasis M22 page) — osmotic dehydration of tissue.

VM-1 / CI-VM-1 (Cryonics Institute)

Cheaper agent, DMSO + ethylene glycol base (CI VM-1 page). Glass transition Tg ≈ −121 to −123 °C at 60–70% w/w. Requires cooling/warming rate >0.1 °C/min to avoid devitrification. CI perfuses at ~−7 °C.

What these agents optimize for

CPAs minimize ice formation during cooling. Their figure of merit is: - Low toxicity at the concentration needed to vitrify. - Penetration kinetics fast enough that whole-organ loading is possible before tissue time-toxicity kills it.

What they don't optimize for: getting past a microcirculation that has been partially or wholly closed by ischemia. They assume a living-quality vascular bed.

2. "CPAs cannot enter closed capillaries" — is this published?

The short answer: yes, implicitly everywhere, but not as a flashy standalone result. The way this shows up in the literature:

• Fahy & Wowk's organ-vitrification work (Fahy 2015; 21CM publications) treats high-quality perfusion of living organs as the starting condition. A kidney coming off a warm ischemic dog has, by protocol, had <30 min of warm ischemia before cold perfusion. The extension to human cryonics post-legal-death is precisely where the unspoken assumption breaks.
• McKenzie et al. 2024 (PMC11416988): explicitly discusses the problem:

  "Postmortem perfusion faces the critical 'no-reflow phenomenon.' ... Many investigators have reported that perfusion after hours of postmortem delay is still useful to help distribute fixatives ... but how rapidly cerebral perfusion degrades in the postmortem period ... is an open question." And on immersion as fallback: "The depth of fluid penetration is proportional to the square root of time, so inner brain regions will undergo a degree of decomposition before fixative reaches the tissue."

• Robert McIntyre's own talks and writings (LessWrong notes on McIntyre talk) explicitly frame ASC's win as "fix first, perfuse second."
• Nectome's 2026 preprint operationalizes it: initial attempts at 18 min post-mortem perfusion showed "evident cellular damage"; <14 min gave "intact membranes, visible mitochondria, and perfectly preserved synapses." This is direct empirical evidence that even with the world's best perfusion protocol and fixative stack, post-ischemia perfusion fails reproducibly past some threshold — and the threshold is capillary-limited. (Lifeboat summary)

So the Aurelia claim "CPAs don't solve the perfusion problem" is a restatement of the empirical ASC motivation: the reason you need aldehyde fixation is precisely that CPAs alone can't get past degraded microvasculature. The simplest positive statement of it I've found is from McKenzie:

"Aldehyde fixation mitigates structural damage during cryopreservation ... likely involves (a) stabilizing membranes to mitigate damage due to dehydration and osmosis, (b) stabilizing blood vessels to improve cryoprotectant perfusion, and/or (c) increasing the cellular permeability of cryoprotectants." (PMC11416988)

Item (b) is the one. (C1)

3. What BPF evaluation checks for

The Brain Preservation Foundation's Large Mammal Prize criteria (BPF prize rules; Large Mammal announcement):

• Intact pig brain preserved such that "the structure of every neuronal process and every synaptic connection remains intact and traceable using today's electron microscopic (EM) imaging techniques."
• Protocol compatible with terminal-patient hospital setting.
• Preservation for centuries (>100 years).
• Evaluation: block horizontally sectioned at 1 mm intervals, cortex to brainstem, each block face imaged at 5 nm pixel size. "The entire surface should be intact and contain clearly visible vesicles and synaptic densities." ([BPF prize rules]; [search summary via Grok/Frontiers])

ASC won in 2018 (21CM/McIntyre); no one has replicated with a different method.

4. ASC's answer: fix first, then perfuse CPA

The ASC protocol outline (McIntyre & Fahy 2015, PubMed 26408851; Nectome 2026):

1. Blood washout with carrier perfusate (B1 or similar).
2. Glutaraldehyde perfusion. Glutaraldehyde cross-links proteins within minutes and stabilizes lipid membranes. Typical laboratory EM fixative: ~2.5% glutaraldehyde. ASC uses a glutaraldehyde-based fixative delivered via vascular perfusion.
3. Gradual ramp of ethylene glycol over several hours, up to ~65% w/v. Eight-step or continuously-ramped protocol.
4. Cooling to −135 °C (below ethylene glycol's Tg ≈ −130 °C at 65%) — no liquid nitrogen needed, just dry-ice-cooled ethanol or equivalent.
5. Long-term storage at "intermediate temperature storage" (ITS) around −135 °C to −140 °C.

The key insight: glutaraldehyde does not prevent ice; ethylene glycol does. Glutaraldehyde prevents the biological damage from cytotoxic edema, membrane lysis, enzymatic autolysis, and — critically — stabilizes capillaries against deforming during the 9-hour CPA ramp. The 2026 Nectome preprint extends this to pig brains under realistic clinical (post-MAiD) conditions with intact volume-EM-traceable synapses.

Evidence for ultrastructural success:

• Rabbit: (McIntyre & Fahy 2015, Cryobiology; BPF eval). FIB-SEM volumes at 8–16 nm isotropic across cortex, striatum. BPF Small Mammal Prize won 2016.
• Pig: BPF Large Mammal Prize 2018. Whole-brain preservation assessed "beautifully preserved across the entire brain."
• Pig post-MAiD-analog protocol: Nectome 2026 preprint; volume EM shows connectomically traceable whole brains when perfusion started <14 min post-arrest.
• Third-party Nectome evaluation (Andrew Critch, Berkeley EM, 2025): "excellent" preservation with "connectomically traceable" tissue; included stress-test samples held at 60 °C for 12 hours with "near-identical" quality (Nectome LessWrong post).

5. Sparks vs Nectome: is there head-to-head EM data?

Sparks Brain Preservation (formerly Oregon Cryonics / Oregon Brain Preservation). Protocol: chemical fixation with aldehyde (formaldehyde + glutaraldehyde), then storage at −20 °C (no vitrification, no cryoprotectant ramp). Cost: ~one-tenth Nectome. Jordan Sparks's argument (oregoncryo, redirects to oregonbp.com; Sparks history): chemical fixation is "the gold standard for structural preservation in neuroscience research for more than a century" and you can do it cheap.

The absent head-to-head: there is no published comparison, at EM scale, between (a) a brain preserved by Sparks with typical low-cost fieldable perfusion and (b) a brain preserved by Nectome with careful perfusion + CPA + vitrification, evaluated by the same independent lab. Sparks has posted some of their own EM results on the Oregon Cryonics / Sparks website, but these are not peer-reviewed, and the brains involved are typically post-mortem human with long ischemic intervals; they are not directly comparable to Nectome's sub-14-minute pig preparations. (C4)

What McKenzie et al. 2024 do say:

"Fluid preservation, which relies on aldehyde fixation, appears to be a cost-effective method." "The use of ASC to preserve an intact pig brain was judged to have met the Brain Preservation Prize's requirement, but this same level of whole connectome preservation quality has not yet been demonstrated in a human brain."

The McKenzie paper's framing: we know ASC works in pig; we don't yet have evidence it works with human, post-mortem, under operational conditions. This is what Nectome's 2026 preprint tries to close.

6. Failure modes specific to CPA perfusion

Working through the mechanical failure tree:

1. Global arterial block (clot, air embolism, dissection). A silent carotid block leaves external-carotid tissue (scalp) perfused so the skin looks OK. You'd only catch it on EM, not on gross inspection. Aurelia called this out specifically. This is also a reason pressure-only quality monitoring is unreliable: the pump happily pushes to the non-blocked side.
2. Capillary-level no-reflow (the main one): ischemic capillary constriction + swelling + neutrophil plugging → CPA can't reach the tissue the capillaries feed. See 01-hemodynamics-and-capillaries.md.
3. Osmotic capillary rupture: high-osmolarity CPA draws water out of tissue, but if done too fast, gradients burst capillary walls. Slow ramps minimize this.
4. Overpressure barotrauma: pushing hard to overcome no-reflow ruptures small vessels; micro-hemorrhages visible on EM.
5. Osmotic brain shrinkage → surface separation: M22 "substantial brain shrinking" can pull the brain away from arachnoid/pial attachments, rupturing bridging veins.
6. Post-vitrification fracturing: thermal stress during cooling cracks the brain. Intermediate temperature storage (ITS) around −140 °C rather than liquid-N₂ −196 °C is the current partial mitigation.
7. Toxicity to already-perfused tissue: M22 held at terminal concentration for too long causes cellular damage even when no ice forms (Fahy 2015 review). Cold storage immediately post-perfusion is essential.

None of #1–#3 are solved by "better CPA." Better CPA helps with #6/#7. This is the key architectural observation: the bottleneck is in the delivery physics, not the chemistry.

7. Summary

• M22 and VM-1 are both well-characterized and close to the limit of what cryobiology can offer for penetrating vitrification.
• The unsolved problem is delivery: even optimal CPA is useless if the capillaries are closed.
• ASC's genius is that glutaraldehyde fixation happens before the long CPA ramp, while vessels are still open, stabilizing both cells and vessels against the subsequent high-osmolarity load.
• The BPF prize criterion ("5 nm pixel EM, every synapse traceable") is a real quality bar; ASC has cleared it at pig scale, others have not.
• Sparks's cost argument is a legitimate alternative philosophy but there is no published head-to-head EM comparison against Nectome-grade preservation on matched samples. The quality question is genuinely open; the cost question is not.